Yifan Xie

Profile

Project title:

Film score composer AI assistant: generating expressive mockups (in collaboration with Spitfire Audio)

Abstract:

This research aims to create an AI system, to help composers create orchestral mockups for film scores. The whole process of creating orchestral mockups is complicated; this research assumes that there are already existing orchestral scores composed by the composer without human performance expressions (Human performance expressions mean how performers or conductors interpret scores, how they control detailed musical expressions like dynamics, articulation, modulation, timing, etc.), and focuses on adding detailed musical expressions (aiming to simulate human to interpret scores) to the computer renditions of the plain scores, aiming to make the music more expressive and realistic. The research will use the digital audio workstation to create music. What this research aims to do is to develop AI methods to automatically add control parameters of detailed musical expressions like CC1 and CC11 to the pre-defined pitches, durations, and tracks.

C4DM theme affiliation:

Music Informatics

Research

Research Interests:

Multimodal Learning, Machine Listening